Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

PHYBench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Video Question AnsweringPhyBench
Avg Score0.7933
12
Physical ReasoningPHYBench
Pass@1 Accuracy5.3
12
KnowledgePHYBench
Score30.06
10
ReasoningPHYBENCH
PHYBench Score26.26
9
Hypothetical Instruction-Based Image EditingPhyBench
Instruction Alignment75.71
7
Physics ReasoningPhyBench
pass@1626
6
Scientific ReasoningPHYBench
Accuracy30.34
5
Showing 7 of 7 rows